Writing a SOAP record that still works months later

A clinical record gets reread in three predictable situations, and they are the three in which you cannot be there to explain: a colleague opens it during a night shift, the patient is referred to an external centre, a complaint requires reconstructing what was known at the time. In all three, the value of the note depends on one thing only: being able to tell what was observed from what was inferred.

This distinction is not formality. If the word pancreatitis appears in the record without any statement of what it rests on, the reader cannot tell whether it was a diagnosis, a suspicion or simply the hypothesis the consultation started from. Six months later that word carries the weight of a confirmed diagnosis and steers decisions nobody questions. The SOAP format, when used for what it is, prevents exactly this drift: the subjective stays reported, the objective stays measured, the assessment stays a declared interpretation, and the plan stays a set of actions with a deadline.

The editor presents the four fields in order and shows a completion bar that advances as they are filled. The bar measures coverage, not quality: its only job is to stop a note being left partial at the end of the day.

In the objective field the difference between a weak note and a solid one lies in the numbers. Recording that the patient is tachycardic says far less than recording the measured rate, because at the next visit comparison is only possible against a value. The ranges below are a reading reference, not a substitute for clinical judgement on the individual patient.

1. 1

   Open a new note before the consultation starts

   Create the note at the beginning, not at the end. The editor saves as you type and shows the save state next to the title, so there is no reason to postpone: what you write during the consultation is more accurate than what you reconstruct three patients later.

2. 2

   Fill in patient name and identifier

   Name and identifier are required and are what make the note findable. If your practice numbers its visits, fill in the visit identifier too: it is the field that separates two close appointments for the same patient without having to read the text.

3. 3

   Pick the template matching the reason for presentation

   The template selector preloads a ready made field structure. Choosing the right template at the start reduces omissions, because the items usually forgotten are already written and only need filling in or explicitly deleting.

4. 4

   Write the subjective while the owner is talking

   Report the account as attributed speech, in the words used, and date it. A line such as reported inappetence for three days is verifiable; a line such as not been himself lately is not. If the owner cannot answer a question, write that down: a declared missing item is worth more than an invented one.

5. 5

   Record the objective as you examine, not from memory

   Enter numbers as soon as you obtain them and describe findings system by system, including the negative ones that rule something out. A physical examination showing only abnormal findings does not tell the reader whether the rest was checked or simply not written down.

6. 6

   In the assessment list problems, not impressions

   Open with a numbered list of active problems. For each one give the hypotheses in order of likelihood and then the elements collected above that support them and those that weaken them. This is the point where it becomes visible whether reasoning happened or was skipped.

7. 7

   In the plan write actions with a deadline

   Every plan item needs a subject, an action and a time: which test, who requests it, by when. Always add the criteria for immediate return, that is the conditions under which the owner must call back without waiting for the scheduled recheck.

8. 8

   Check the completion bar before closing

   The editor flags a note that is still partial. Partial notes at the end of a shift are the main cause of unreadable records: if a field stays empty for a specific reason, write that reason inside the field instead of leaving it blank.

9. 9

   Save, close and reread the next day

   Use the explicit save on closing and confirm the state shows the note as saved. The following day reopen one note at random from those you wrote and apply the thirty second test: it is the only quality control that survives over time without organising meetings.

The assessment is the section filled last, with the patient already in the waiting room and the next one due. It is also the only one that makes everything else usable. A quick way not to leave it empty is to complete it with three fixed lines for each active problem.

1. The problem, stated at the level of certainty you actually have: a clinical sign if that is all you know, a syndrome if you have a picture, a diagnosis only if you hold the confirming datum.
2. Hypotheses in order of likelihood, three at most: a long list does not help the reader and does not reflect how reasoning really works.
3. Supporting and opposing elements, taken from what you already wrote in the objective field: if an element does not appear above, either add it above or do not use it here.
4. Comparison with the previous visit in one line: improved, stable or worse against which measured parameter.
5. What would change the assessment, that is which expected result would make you abandon the leading hypothesis.

The templates area holds predefined record structures, searchable by name, and lets you create a preformatted note by choosing the template and naming the patient. The real benefit is not typing speed, it is uniformity: if the whole team starts from the same structure, comparing two visits by the same patient becomes immediate even when different people carried them out.

• Choose the template before you start writing, not halfway: swapping it later forces you to redo fields already completed.
• Explicitly delete template items that do not apply instead of leaving them blank: a blank item does not separate not assessed from not present.
• Do not reuse sentences written for another patient without rereading them: a copied description that does not fit is worse than an empty field.
• Update templates when an item is consistently useless or consistently missing, and tell the team you did.
• Keep a shorter template for rechecks: a recheck note repeating the whole structure of a first visit gets filled badly by everyone.

A monthly review of a small sample of notes, run against written criteria, produces more improvement than any general reminder. Five randomly drawn records and the table below are enough, completed by someone other than the author.